August ii, 1921] 



NATURE 



759 



that the kinetic theory of gases allows us, from 

 measurements of the viscosity or the heat conductivity 

 of a gas, to calculate the mean distance of the centres 

 of two molecules of the gas apart when the molecules 

 are in contact, that the constant h of Van der Waals 

 furnishes another estimate of the distance, and that 

 the two estimates agree in giving about 2 x 10-' cm. 

 for the mean radius of hydrogen and helium molecules 

 and about 3 x 10- ' cm. for the mean radius of the 

 molecules of argon, nitrogen, oxygen, carbon dioxide, 

 and other gases. With these figures as a basis, X-ray 

 crystal analysis then gives the relative positions of the 

 atoms in the molecule of the material analysed. So 

 far the most careful analysis of crystals of potassium 

 chloride has, however, failed to reveal any structure 

 corresponding to the KCl molecule. Each K atom 

 is surrounded by six CI atoms at equal distances from 

 it. For chlorides the distances vary with the metal in 

 the molecule, are large, —3 to 5x10-' cm., for the 

 first elements of a "period," and decrease to a limit 

 1-3 to 27 X 10-' for the last elements. 



Many methods of harmonic analysis have been 

 given of recent years. We need mention only the 

 methods of Perry, Silvanus Thompson, and Russell. 

 The question has now come prominently forward in 

 connection with the disturbances induced in telephone 

 and radio stations bv the harmonics in the currents 



carried by overhead power lines. The power station 

 engineer wants the manufacturer to guarantee that 

 the electric generator he purchases from him shall 

 give a pure sine-shaped wave of electromotive force. 

 As it is impossible to make the machine give an 

 absolutely pure sine wave, limits have to be fixed on 

 the magnitudes of the amplitudes of the harmonics in 

 the wave. Hence harmonic analysis is a necessity. 

 In the Journal of the Institution of Electrical En- 

 gineers (vol. lix., p. 491) Mr. A. E. Clayton gives a 

 resume of the ordinary methods and two schedules for 

 "harmonic analysis " by means of selected ordinates. 

 One goes to the 25th harmonic and the other to the 

 13th. In the one case the assumption is made that 

 no harmonic higher than the 25th is present, and in 

 the other that there is none higher than the 13th. 

 Seeing that in actual electromotive-force w-aves there 

 is an infinite number of harmonics present, and as 

 only a limited number of ordinates are drawn, we 

 should have little confidence in results obtained by a 

 " schedule." 



The Cambridge University Press will publish 

 shortly "The Calendar," by A. Philip, the purpose 

 of which is to provide a concise and popular summary 

 of the history and construction of the Gregorian 

 calendar, with special reference to the reform of the 

 calendar and the fixing of the Easter date. 



Our Astronomical Column. 



Bright Objfxt near the Sun.- — Prof. Campbell, 

 Director of the Lick Observatory, reports by telegraph 

 an object brighter than Venus that was seen on 

 August 7 3° east of the sun and 1° south. The mes- 

 sage states that there is no doubt of the object being 

 a celestial object. It is either a comet or a nova. 

 The former appears more probable, owing to the dis- 

 tance from the Galaxy, where most novae appear. 



Variable Stars. — The Bruce 24-in. photographic 

 telescope at Arequipa has been used for taking 

 spectrograms of the Large Magellanic Cloud, in 

 which Miss Leavitt some years ago detected several 

 variable stars (Harv. Ann., vol. Ix., No. iv.). Miss 

 Cannon, in Harv. Bull. No. 754, gives the spectral 

 type and magnitude range of eight of them as 

 follows :— No. 884, Mc, ii-4m. to i5-5m. ; No. 900, 

 M, i2-2m. to i3-6m. ; No. 2257, K5, i2-4m. to i3-2m, ; 

 No. 2435, K5, io-8m. to ii-7m. ; No. 2447, K5, 120m. 

 to i2-8m. ; No. 2622, K5, i3-2m. to 140m. ; No. 2822, 

 Mc, 9-8m. to io-6m. ; and No. 2882, Mb, iiom. to 

 i3-6m. The numbers are from Harv. Ann., vol. Ix. 

 It is satisfactory to find that such faint stars are 

 within the reach of spectroscopic analysis. 



Mr. Stanley Williams contributed a paper to Monthly 

 Notices, R.A.S., vol. Ixxxi., p. 332, on the star 

 B.D. + 44-994°, which he announced as a peculiar 

 variable, possibly of the Cepheid type. Miss Cannon 

 gives its spectral type as Ma ; and Miss Leavitt has 

 identified 150 images of the star on plates taken 

 during the last twenty years. Its normal photo- 

 graphic magnitude is io-5m., but on seven dates it 

 was io-2m. ; it is very red (Harv. Bull., No. 754). 



C. Hoffmeister, director of Sonneberg Observatory, 

 noted on May 30 last, while observing Reid's comet, 

 an 8th magnitude star that is not in the B.D. Its 

 position for 18550 is 7h. 570m., N. 58° 14', and it is 

 shown on the Harvard plates. Prof. Kiistner has 



NO. 2702, VOL. 107] 



examined the original observations of the B.D., and 

 finds that a star of 9-5m. was observed in the place 

 on February 19, 1858, but not seen again, so it is 

 probably variable {Astr. Nach., Circ. 22). In the 

 same circular H. Fuss announces that B.D. +42-3351°, 

 7-5m., has the large proper motion of + 0065s., 

 +o-i6". 



Mr. Flint's Parallax Observations. — Publications 

 of Washburn Observatory, vol. xiii., part i, contains 

 the details of the series of meridian observations for 

 stellar parallax made at Washburn between 1898 and 

 1905 with the Repsold meridian circle of 12-2 cm. 

 aperture, fitted with a travelling-wire micrometer. 



The programme extended from declination —35° to 

 + 90°, and embraced stars from magnitude 1-5 to 2-5, 

 with some binaries and stars of sensible proper motion. 

 A screen with thin metal . slats rotating about their 

 axes like the laths of a Venetian blind was used to 

 equalise magnitudes, 70m. being made the standard. 

 Two comparison stars, one preceding, the other fol- 

 lowing, the parallax star, were used in each case. 



The mean probable errors of a single observation of 

 unit-weight and of the final parallax of each star are 

 0-214" and 0-031" respectively. The last quantity is of 

 about three times the size of the probable error in 

 the best recent photographic determinations, showing 

 that the meridian method cannot compete seriously 

 with the photographic. Still, the experiment was 

 well worth making, and the research will occupy a 

 place in the historv of the subject, so that it is well 

 to have the details published. The list of parallaxes 

 contains 124 stars, of which the deduced relative 

 parallax is negative in thirty-five cases. The values 

 for Algol and Castor, 0-122" and 0167", are about three 

 times the accepted values, while that of Altair, 0071", 

 is only about one-third of it ; but in many cases there 

 is better agreement. 



